Juan Carlos Munoz Morano

How about an example file so we can see what you're working with?

Juan Carlos Munoz Morano said:

The practical case would be a huge file that has a lot of repetitions.

I wouldn't use unknown segments, although you could.  Probably easier to simply filter on them, handle them all in one go, change status to translated and then lock them or hide them.

Hi Paul,

Thanks for your prompt answer!

Attached you have the test files that I used for this. I have already extracted the Unknown Segments file, translated it and populated the TM with it. Please let me know with your findings!

The reason why I use the Unknown Segments file:

• When working with hundreds of even thousands of files this option provides a lot of efficiency, since you only have to work with one file.
• When extracting Unknown Segments the file size also decreases significantly (also for package creation)
• Studio also works more smoothly when dealing with just one file.

Another option would be running two rounds of pre-translation of the source files. One with an empty TM so all the tokens are auto-localized and another one with the translated Unknown Segments file. However, I still face the same issue here, since not all the tokens are auto-localized even with the empty TM.

The option of not recognizing tokens in the TM used to extract the Unknown Segments has been also considered, however,  the word count increases significantly since all the tokens would be included as common text and we would losing the opportunity of this auto-localization that Studio offers.

What am I missing? Maybe some settings that I did not pay attention to?

Thank you very much in advance

Carlos

SDL Unknown Segments Test files.zip

Hi Paul,

Thanks for your prompt answer!

Attached you have the test files that I used for this. I have already extracted the Unknown Segments file, translated it and populated the TM with it. Please let me know with your findings!

The reason why I use the Unknown Segments file:

• When working with hundreds of even thousands of files this option provides a lot of efficiency, since you only have to work with one file.
• When extracting Unknown Segments the file size also decreases significantly (also for package creation)
• Studio also works more smoothly when dealing with just one file.

Another option would be running two rounds of pre-translation of the source files. One with an empty TM so all the tokens are auto-localized and another one with the translated Unknown Segments file. However, I still face the same issue here, since not all the tokens are auto-localized even with the empty TM.

The option of not recognizing tokens in the TM used to extract the Unknown Segments has been also considered, however,  the word count increases significantly since all the tokens would be included as common text and we would losing the opportunity of this auto-localization that Studio offers.

What am I missing? Maybe some settings that I did not pay attention to?

Thank you very much in advance

Carlos

SDL Unknown Segments Test files.zip

Juan Carlos Munoz Morano

Thanks for the file. I cleaned your TM a little so I had no numbers, or acronyms, in there and ran a pre-translation.  This is what I get:

The numbers are recognised as expected, but the acronyms are not.  The reason for this is that we can't be sure whether an acronym remains the same in pre-translation or not.  So it's only available for interactive translation where the user can make the choice.

I guess if you think this should be possible as an option you should raise this here and see whether you get any support:

http://ideas.sdl.com

Hi Paul,

Thank you very much for your feedback.

The conclusion is then that, when using this Unknown Segments approach, these acronyms should not be recognized in the TM so that they are fully auto-translated once the Unknown Segments file is translated (meaning this the mentioned increase of word to be translated) or manually fixing them (with regex, for example) since they should be just copied to target. Meaning this that there is no other way to automatically localized them.

Thank you for your time

Kind regards,

Carlos

Juan Carlos Munoz Morano

Juan Carlos Munoz Morano said:

Meaning this that there is no other way to automatically localized them.

Also... once in your TM they'll be picked up so I guess over time you'll start to incorporate the sort of acronyms you handle and then they'll be pre-translated in future files.